One huge blind spot in using panel data over a 34 year period is that a lot of the people you are initially looking at are retired, and so do not typically pay as much tax or otherwise need to make the same amount of income to have the same standard of living.

The article is very poorly framed and argued.